English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Performance variability is an important measure for a reliable highperformance computing (HPC) system. Performance variability is affected bycomplicated interactions between numerous factors, such as CPU frequency, thenumber of input/output (IO) threads, and the IO scheduler. In this paper, wefocus on HPC IO variability. The prediction of HPC variability is a challengingproblem in the engineering of HPC systems and there is little statistical workon this problem to date. Although there are many methods available in thecomputer experiment literature, the applicability of existing methods to HPCperformance variability needs investigation, especially, when the objective isto predict performance variability both in interpolation and extrapolationsettings. A data analytic framework is developed to model data collected fromlarge-scale experiments. Various promising methods are used to build predictivemodels for the variability of HPC systems. We evaluate the performance of themethods by measuring prediction accuracy at previously unseen systemconfigurations. We also discuss a methodology for optimizing systemconfigurations that uses the estimated variability map. The findings frommethod comparisons and developed tool sets in this paper yield new insightsinto existing statistical methods and can be beneficial for the practice of HPCvariability management. This paper has supplementary materials online.

Prediction of High-Performance Computing Input/Output Variability and Its Application to Optimization for System Configurations